Anomalies in Trace Metal and Rare-Earth Loads below a Waste-Water Treatment Plant

The changes in chemical loads were examined for 54 inorganic elements and compounds in a 5.4-km reach of Boulder Creek, Colorado downstream of a waste water treatment plant (WWTP) outfall. Elements were partitioned into three categories: those showing a decrease in loading downstream, those showing an increase, and those which were conservative, at least over the length of the study reach. Dissolved loads which declined - generally indicative of in-stream loss via precipitation or sorption - were typically rapid (occurring largely before the first sampling site, 2.3 km downstream); elements showing this behavior were Bi, Cr, Cs, Ga, Ge, Hg, Se and Sn. These results were as expected before the experiment was performed. However, a large group (28 elements, including all the rare-earth elements, REE, except Gd) exhibited dissolved load increases indicating in-stream gains. These gains may be due to particulate matter dissolving or disaggregating, or that desorption is occurring below the WWTP. As with the in-stream loss group, the processes tended to be rapid, typically occurring before the first sampling site. Whole-water samples collected concurrently also had a large group of elements which showed an increase in load downstream of the WWTP. Among these were most of the group which had increases in the dissolved load, including all the REE (except Gd). Because whole-water samples include both dissolved and suspended particulates within them, increases in loads cannot be accounted for by invoking desorption or disaggregation mechanisms; thus, the only source for these increases is from the bed load of the stream. Further, the difference between the whole-water and dissolved loads is a measure of the particulate load, and calculations show that not only did the dissolved and whole-water loads increase, but so did the particulate loads. This implies that at the time of sampling the bed sediment was supplying a significant contribution to the suspended load. In general, it seems untenable as a hypothesis to suppose that the stream bed material can permanently supply the source of the in-stream load increases of a large group of inorganic elements. We propose that the anomalous increase in loads was more a function of the time of sampling (both diurnally and seasonally) and that sampling at different times of day or different seasons during the year would give contradictory results to those seen here. If this is so, inorganic loading studies must include multiple sampling both over the course of a day and during different seasons and flow regimes.